An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

1

Predicting the drying properties of sludge based on hydrothermal

treatment under subcritical conditions

Mikko Mäkelä*a, Laurent Fraikinb, Angélique Léonardb, Verónica Benaventec, Andrés Fullanac

a Swedish University of Agricultural Sciences, Department of Forest Biomaterials and Technology,

Skogsmarksgränd, 90183 Umeå, Sweden

Current address: Tokyo Institute of Technology, Department of Environmental Science and

Technology, G5-8, 4259 Nagatsuta-cho, Midori-ku, Yokohama 226-8502, Japan

b University of Liège, Department of Chemical Engineering, Agora, B6c, Allée du 6 Août, 4000

Liège, Belgium (L. Fraikin: laurent.fraikin@ulg.ac.be, A. Léonard: a.leonard@ulg.ac.be)

c University of Alicante, Department of Chemical Engineering, P.O. Box 99, 03080 Alicante, Spain

(V. Benavente: veronica.benavente@ua.es, A. Fullana: andres.fullana@ua.es)

*corresponding author, tel.: +81 (0)45 924 5507, fax: +81 (0)45 924 5507, email:

makela.m.aa@m.titech.ac.jp, mikko.makela@slu.se

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

2

Abstract

The effects of hydrothermal treatment on the drying properties of sludge were determined. Sludge

was hydrothermally treated at 180-260 °C for 0.5-5 h using NaOH and HCl as additives to influence

reaction conditions. Untreated sludge and attained hydrochar samples were then dried under identical

conditions with a laboratory microdryer and an X-ray microtomograph was used to follow changes in

sample dimensions. The effective moisture diffusivities of sludge and hydrochar samples were

determined and the effect of process conditions on respective mean diffusivities evaluated using

multiple linear regression. Based on the results the drying time of untreated sludge decreased from

approximately 80 minutes to 37-59 minutes for sludge hydrochar. Drying of untreated sludge was

governed by the falling rate period where drying flux decreased continuously as a function of sludge

moisture content due to heat and mass transfer limitations and sample shrinkage. Hydrothermal

treatment increased the drying flux of sludge hydrochar and decreased the effect of internal heat and

mass transfer limitations and sample shrinkage especially at higher treatment temperatures. The

determined effective moisture diffusivities of sludge and hydrochar increased as a function of

decreasing moisture content and the mean diffusivity of untreated sludge (8.56 ∙ 10-9 m2 s-1) and

sludge hydrochar (12.7-27.5 ∙ 10-9 m2 s-1) were found statistically different. The attained regression

model indicated that treatment temperature governed the mean diffusivity of hydrochar, as the effects

of NaOH and HCl were statistically insignificant. The attained results enabled prediction of sludge

drying properties through mean moisture diffusivity based on hydrothermal treatment conditions.

Keywords: Biosolids; Moisture diffusivity; Hydrochar; Hydrothermal carbonization; Shrinkage; X-

ray microtomography

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

3

1. Introduction

Handling of sludge residues generated by biological and chemical wastewater treatment is becoming

ever more challenging due to rapid urbanization and increasing efficiency requirements for municipal

and industrial wastewater treatment plants. Sludge residues are widely complex materials due to a

variety of structural components such as extracellular polymeric structures, filamentous bacteria,

cationic salts and the potential presence of pollutant precursors, e.g., proteins and fats, pathogens,

parasites, trace metals, polychlorinated biphenyls, dioxins and other slowly decomposable compounds

(Vaxelaire and Cézac, 2004; Yoshikawa and Prawisudha, 2014b). Currently the most common sludge

handling methods include incineration, composting, use in agriculture or disposal in landfills

(Mahmood and Elliott, 2006; Mowla et al., 2013), although current trends in European regulation are

increasingly hindering landfill deposition of organic substances. As sludge is an inherently wet

material decreasing associated handling, storage and transportation costs generally requires active

drying especially for smaller-scale treatment plants. However current means of mechanical

dewatering suffer from difficulties in removing intracellular and chemically bound water from the

polymeric matrix of sludge suspensions (Mowla et al., 2013; Stoica et al., 2009).

Hydrothermal treatment performed with water acting as a solvent, a reactant and even a catalyst or

catalyst precursor does not require prior drying of a sludge feedstock and allows simultaneous

elimination of biologically active organisms or compounds (Peterson et al., 2008). Hydrothermal

treatment under subcritical conditions can be used for producing carbonaceous hydrochar with

relatively high yields (Libra et al., 2011) and enables the conversion of non-traditional feedstock such

as municipal and industrial sludge residues. The increasing ion product of water under hydrothermal

conditions typically favours reactions which are catalysed by acids or bases and is generally

understood to proceed via a network of hydrolysis, dehydration, decarboxylation, polymerization and

aromatization of biomass components (Kruse et al., 2013; Sevilla and Fuertes, 2009). As

hydrothermal treatment is ideally operated under saturated steam pressure the latent heat requirement

of evaporation can also be avoided although post-separation of solid and liquid phases is still required.

Irrespective of potential hydrochar applications such as direct solid fuel replacement, soil

amelioration or metal/metalloid adsorption (Alatalo et al., 2013; Libra et al., 2011; Titirici and

Antonietti, 2010), elimination of hydroxyl and carboxyl groups followed by subsequent aromatization

can enhance the drying properties of sludge due to increased hydrophobicity and sludge cell breakage

(Wang et al., 2014; Yoshikawa and Prawisudha, 2014a; Zhao et al., 2014b).

The drying properties of a material are conveniently characterized through the drying curve where

drying rate is described as a function of drying time under specific drying conditions. The attained

drying curve allows the identification of controlling mechanisms such as moisture evaporation from

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

4

saturated or unsaturated surfaces or diffusion within a material (Mujumdar, 2007). In addition,

effective moisture diffusivity can be used for compiling empiric drying data to a single parameter

describing moisture transfer independent of the actual mechanisms involved (Gómez-de la Cruz et al.,

2015). Although diffusivity coefficients are often determined through a simplification of Fick’s

second law of diffusion, reliable estimations for non-rigid materials can only be attained if material

shrinkage is taken into account (Bennamoun et al., 2013). Non-extrusive imaging techniques such as

X-ray microtomography have been shown to provide detailed quantitative information on material

shrinkage and thus enable a better understanding of the relationships between drying properties and

the evolution of size and shape (Léonard et al., 2004; Léonard et al., 2008).

For reliably determining the effect of hydrothermal treatment on the drying properties of sludge, we

illustrate the use of X-ray microtomography for monitoring sample shrinkage during drying of sludge

and hydrochar samples. Respective effective moisture diffusivities affected by shrinkage were then

determined by applying recent developments on the interpretation of experimental drying data.

Finally, the effect and statistical significance of treatment conditions such temperature, retention time

and additive quality on the mean moisture diffusivity of hydrochar were determined using multiple

linear regression. The attained results help in understanding the effect of process conditions on the

drying properties of hydrothermally treated sludge and controlling these properties based on treatment

conditions.

2. Material and methods

2.1 Sampling and sample preparation Sludge samples were attained from a Swedish pulp and paper mill using virgin sulphate and recycled

fibre pulp for the production of unbleached kraft/euroliner for corrugated cardboard. Mill effluents

were treated by primary gravitational settling followed by biological activated sludge treatment.

Approximately 300 kg of mixed sludge (containing 60% of primary sludge and 40% of biosludge)

was sampled after primary sludge and surplus biosludge from secondary sedimentation had been

mixed and dewatered to approximately 27% dry solids (2.7 kg H2O kg-1 db, dry basis) using a belt

filter and a centrifuge at the mill. The 300 kg sample was coned and quartered (Gerlach et al., 2002)

to a representative 10 kg subsample which was stored in +4 °C during the experiments.

2.2 Hydrothermal treatment Hydrothermal experiments were performed with a 1 L non-stirred stainless steel reactor (Amar

Equipments PVT Ltd., Mumbai, India) illustrated Fig. 1a. A constant 300 g mass of sampled sludge

was thoroughly mixed with 75 mL of additive and loaded into the reactor. The reactor was heated to

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

5

reaction temperature using a 1.5 kW electric heating resistance and an additional heating plate placed

under the reactor. The 1.5 kW heating resistance was PID controlled as the additional heating plate

was set to reaction temperature. Reactor pressure was indicated by a pressure gauge and was

approximately equivalent to saturated vapour pressure of water under respective reaction temperatures

(i.e., 1-5 MPa). As the isothermal retention time was complete, the reactor was cooled with

pressurized air and the gases released into a fume hood. The solid and liquid phases were

subsequently separated by vacuum filtration through a grade 413 VWR® filter paper (VWR

International LLC, Radnor, PA, USA).

Fig. 1: Schematic presentations of a) the hydrothermal reactor and b) the microdryer.

The experiments were conducted according to an experimental design including reaction temperature

(180-260 °C) and log10 transformed retention time (0.5-5 h) as continuous controlled variables. In

addition, additive type was included as a discrete variable to influence reaction conditions by mixing

75 mL NaOH (0.01 N, pH 12.1) or HCl (0.01 N, pH 2.5) with the feed material prior to loading into

the reactor. Deionized H2O was used as a control (conductivity <6 mS cm-1) and both NaOH and HCl

used for making the solutions were reagent grade. NaOH was chosen as it is commonly used in the

chemical recovery cycles of pulp mills and the inclusion of HCl enabled evaluating the effect of a

wider pH range. Additive concentrations were chosen based on Lu et al. (2014) and the final solid

load adjusted to remain in the range applicable to dewatered sludge within the pulp and paper

industry. As the design included a discrete variable it was constructed to allow the use of dummy

variables in linear regression and response surface methodology (Myers et al., 2009). The final design

(see Table 1) was composed of 15 individual experiments.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

6

Table 1: Hydrothermal treatment conditions and mean moisture diffusivities of hydrochar.

Exp. n:o

Reaction temperature (°C)

Retention time Additive Determined mean effective diffusivity (10-9 m2 s-1)a

h log h 1 180 0.50 -0.3 NaOH 12.7 2 260 0.50 -0.3 NaOH 17.9 3 180 5.00 0.699 NaOH 13.9 4 260 5.00 0.699 NaOH 27.5 5 180 1.58 0.1995 NaOH 13.6 6 220 5.00 0.699 NaOH 18.1 7 180 0.50 -0.3 H2O 14.8 8 260 0.50 -0.3 H2O 18.2 9 180 5.00 0.699 H2O 16.4 10 260 5.00 0.699 H2O 23.8 11 180 0.50 -0.3 HCl 13.2 12 260 0.50 -0.3 HCl 16.7 13 180 5.00 0.699 HCl 17.0 14 260 5.00 0.699 HCl 18.9 15 220 1.58 0.1995 HCl 14.7 a Mean effective moisture diffusivity of untreated sludge 8.56 ∙ 10-9 m2 s-1

2.3 Drying experiments and X-ray microtomography Prior to drying the individual sludge and hydrochar samples were shaped to ensure comparability of

initial stress states. The samples were compressed for 1 minute under 50 N in a cylindrical

compression cell (∅ 20 mm) consisting of a movable piston and a closed grid that allowed water to be

removed. The obtained cylindrical samples were then cut to lengths equivalent to 4.10 ± 0.041 g in

sample mass. The corresponding sample volumes were hence dependent on the specific properties of

sludge or hydrochar.

Drying experiments were performed with a laboratory micro-dryer by following sample mass loss

under constant and reproducible drying conditions, Fig. 1b (Léonard et al., 2002). The prepared

samples were inserted to a drying chamber (cross-section 41 × 46 mm) on a grid suspended under a

precision scale to allow convective drying on the entire external surface. The mass of the samples

were then recorded every 5 seconds under an air velocity of 1.5 m s-1, a temperature of 105 °C and an

absolute humidity of 0.007 kg H2O kg-1 dry air. Drying was continued until no further changes in

sample mass were observed and the dried samples placed into an oven at 105 °C to determine

respective dry solids contents according to standard methods. The mass loss signals were then

preprocessed by excluding data points above 95% dry solids (0.053 kg H2O kg-1 db) for removing

interferences at low moisture contents.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

7

External surface areas and sample volumes were calculated based on image analysis by X-ray

microtomography. Each sample was scanned before and after drying (Fig. 2) with a 1074 portable X-

ray micro-CT scanner (Skyscan, Kontich, Belgium) for 3 minutes resulting in 105 projections. Each

sample was tilted around 180° recording one projection every 1.7° for minimization of acquisition

time. Sample masses were also measured before and after tomography for linking external surface

areas and sample volumes to respective moisture contents (Fraikin, 2012). Sample shrinkage was

assumed linear in respect to sample moisture contents (Léonard et al., 2004; Li et al., 2014).

Fig. 2: Example X-ray images of untreated mixed sludge a) before and b) after drying.

3. Calculations

The drying fluxes of sludge and hydrochar samples were calculated by correcting the recorded drying

data with the external surface area of each sample from X-ray imaging. In addition, respective sample

volumes were calculated. Raw drying data was also interpreted based on Fick’s second law of

diffusion on the hypothesis that moisture transfer was proportional to the concentration gradient of

desorption (Bennamoun et al., 2013):

!"!"= 𝐷!"",! ∙

!!!!!!

+ 𝐷!"",! ∙!!!!!!

+ 𝐷!"",! ∙!!!!!!

(1)

where X denotes moisture content (kg H2O kg-1 db), t drying time (s), Deff the effective diffusivity (m2

s-1) and x, y and z the spatial dimensions of moisture transport (m). For short cylinders approaching a

one-dimensional plane sheet with uniform initial moisture distribution Eq. (1) can be expressed as

(Crank, 1975; Pacheco-Aguirre et al., 2014):

𝜙 = !(!)!!!!!!!!

= !!!

!(!!!!)!

exp  (− (!!!!)!!!!!""!!!

!!!! ) (2)

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

8

where 𝜙 describes dimensionless moisture ratio, X(t) moisture content (kg H2O kg-1 db) at time t, Xe

equilibrium moisture content approaching zero on longer drying times, X0 the initial moisture content,

n a positive integer and L sample height (m), i.e. sheet thickness. For considerable drying times Eq.

(2) can be simplified by considering only the first term of the series, which becomes (Crank, 1975):

!!"ln  (! !

!!) = − !!""!!

!! (3)

Effective moisture diffusivity can then be determined through:

𝐷!"" = − !!!

!! (4)

where k is attained by derivating a polynomial function fitted to the experimental ln ! !!!

data.

The effects of controlled design variables, i.e. reaction temperature, retention time and additive

quality, on the effective diffusivities of hydrochar were modelled through a multiple linear regression

equation:

𝐲 = 𝐗𝐛 + 𝐞 (5)

where y represents a vector of determined mean diffusivities, X a coded and range-scaled design

matrix including individual experiments as rows and experimental conditions as the respective

columns, b a vector of model coefficients and e a residual vector. Model coefficients were determined

by minimizing the sum of squares of model residuals through the least-squares fit:

𝐛 = (𝐗!𝐗)!𝟏𝐗′𝐲′ (6)

Individual model coefficients were refined by F-testing the variance explained by a coefficient against

the variance of model residuals calculated as the difference between determined and predicted

diffusivities:

𝐞 = 𝐲 − 𝐲 (7)

where 𝐲 is a vector of predicted mean diffusivities:

𝐲 = 𝐗𝐛 (8)

The effect of additive quality included as a discrete variable was determined through the use of two

dummy variables c1 and c2 in the design matrix X where:

𝑐! = !  !"  !"#$  !"  !"#  !"#$%&'&  !"#"!!  !"#!$%!&! (9)

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

9

𝑐! = !  !"  !"#  !"  !"#  !"#$%&'&  !"#"!!  !"#!$%!&! (10)

The variation of determined diffusivities explained by the model was calculated through the R2 value:

𝑅! = 1 − !!!"#!!!"!

(11)

where SSres denotes the sum of squares of model residuals and SStot the total sum of squares of y

around the mean.

4. Results

A drying time of approximately 80 minutes was required to reach 95% dry solids in the untreated

sludge sample and respectively decreased to a range 37-59 minutes for hydrothermally treated

hydrochar. Time required for drying was accompanied by considerable sample shrinkage especially in

the case of untreated sludge and was inversely correlated with hydrothermal treatment temperature (p

< 0.01) and retention time (p < 0.01). The external surface area and volume of sludge and hydrochar

samples assuming linear isotropic shrinkage were attained through X-ray tomography and used for

calculating the drying flux as a function of sample moisture content. The results are provided in Fig.

3a and b.

The uncorrected drying data were further interpreted based on Fick’s second law of diffusion in a one-

dimensional plane sheet. Moisture ratio was first determined as a function of drying time (Fig. 3c) and

the respective natural logarithm fitted with a second order polynomial (Fig. 3d). The attained

polynomial coefficients represented a theoretical fit to the experimental data and were used together

with respective sample heights (Fig. 3e) for determining the effective moisture diffusivity of sludge

and hydrochar samples (Fig. 3f). As illustrated in Fig. 3f, the determined moisture diffusivities were

inversely related to respective moisture contents during drying and thus increased towards the end of

the drying experiments. The mean values of determined diffusivities of hydrochar were in the range

12.7-27.5 ∙ 10-9 m2 s-1 (Table 1) and were found statistically different from the moisture diffusivity of

untreated sludge (8.56 ∙ 10-9 m2 s-1) on a p < 0.01 significance level according to a performed t-test.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

10

Fig. 3: a) Drying flux; b) Isotropic volume ratios; c) Moisture ratios (Xt / X0); d) Polynomial fit of ln

(Xt / X0); e) Isotropic dimension ratios; and f) Effective moisture diffusivities (10-9 m2 s-1) of sludge

and hydrochar as a function of moisture ratio. In a) only hydrochar samples produced at 180 and 260

°C and in b)-f) only three samples (untreated sludge and exp. no. 11 and 14, Table 1) and every 50th

data point are shown for clarity.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

11

Coded and range-scaled regression model coefficients for predicting the mean moisture diffusivity of

hydrochar were determined and are illustrated in Fig. S1 (Supplementary Information). Both

hydrothermal treatment temperature (p < 0.01) and transformed retention time (p < 0.05) were

statistically significant in predicting the mean moisture diffusivity of hydrochar according to the

model. No statistically significant relationship was found between mean moisture diffusivity and the

dummy variables used for describing additive quality (p > 0.05). The F-test performed against model

residuals indicated that the attained model was statistically significant on a p < 0.01 significance level

(Table S1). In addition, the calculated R2 value suggested that the acquired model explained a

satisfactory 76% of variation in the determined mean diffusivities of hydrochar samples.

5. Discussion

As illustrated in Fig. 3a, hydrothermal treatment led to considerable differences in the drying behavior

of sludge and hydrochar samples. The drying of untreated sludge began by a short, sharp heating

period where the drying flux increased rapidly due to increase in sample temperature. The constant

drying rate period generally controlled by vapor transfer at the air-solid interface and the rate of

moisture removal could only be seen as a slight increase in drying flux at approximately 2.5 kg H2O

kg-1 db due to simultaneous sample shrinkage (Fig. 3b). On the contrary, drying was mainly governed

by the falling rate period during which the drying flux decreased continuously with a decrease in

sludge moisture content. This period likely began as surface moisture was sufficiently reduced and

moisture was transported to the surface of the solid by capillary forces (Bennamoun et al., 2013;

Mujumdar, 2007). Moisture diffusion caused by concentration gradients within the sample matrix and

heat conduction are generally regarded as the controlling mechanisms (Léonard et al., 2002).

In the case of hydrochar the initial moisture contents were lower due to hydrothermal treatment and

subsequent solid-liquid separation by filtering. Hydrothermal treatment increased the drying flux of

hydrochar and respectively decreased the length of the falling rate period (Fig. 3a) most likely due to

sludge cell breakage (Dawei et al., 2012). In addition, two different phases could be observed during

the falling rate period, the phases being especially distinguishable at approximately 0.5 kg H2O kg-1

db for samples treated at 260 °C. As discussed by Léonard et al. (2002), the first phase of the falling

rate period is normally governed by decreasing drying air humidity at the air-solid interface due to

heat and mass transfer limitations and simultaneous sample shrinkage. During the second phase

sample shrinkage decreases or ceases entirely as the decreasing drying flux is almost entirely caused

by internal heat and mass transfer limitations (Léonard et al., 2002). Our observations on sample

shrinkage in Fig. 3b do not entirely support this description as they assume linear isotropic shrinkage

measured before and after the drying experiments. However, it is more likely that a higher degree of

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

12

shrinkage occurred in the beginning of the drying process during the constant rate period and the first

phase of the falling rate period (Fig. 3a) especially with hydrochar samples where different falling rate

periods could be distinguished. As illustrated in Fig. 3b hydrothermal treatment decreased sample

shrinkage and the effect of heat and mass transfer limitations based on an increased drying flux.

Moisture diffusivity coefficients are often determined based on a simplification of Fick’s second law

of diffusion assuming isotropic diffusion (Pacheco-Aguirre et al., 2014). As illustrated in Fig. 3f, the

determined effective moisture diffusivities of sludge and hydrochar increased as a function of

decreasing moisture content. As described by Fyhr and Kemp (1998), moisture diffusivity in solids is

nearly constant at higher moisture contents but generally reaches a peak when no free moisture is left

in the solid and the only remaining mechanisms for moisture transport are diffusion and convective

vapour flow. However, the peak is generally followed by a notable decrease in diffusivity as the

moisture content of the solid approaches zero. In our data the determined diffusivities turned towards

zero at low moisture contents, but could not be observed in Fig. 3f as the drying data was

preprocessed by excluding data points beyond 95% dry solids. This affected the coefficients attained

by fitting a polynomial function to the experimental data, but ensured comparability of mean moisture

diffusivities between different samples. Recently Gómez-de la Cruz et al. (2015) reported a

modification for the determination of effective diffusivities by using a second order polynomial for

describing moisture ratio as a function of time. The authors also showed increasing effective

diffusivities as a function of decreasing moisture contents and increasing drying temperature for rigid

olive stone. In our case the modification yielded R2 values in the range 0.996-0.999 for sludge and

hydrochar samples and was hence adopted (Fig. 3d). This method also allowed taking sample

shrinkage into account by including a time dependent function for sample height. As illustrated in Fig.

3b and e considerable sample shrinkage occurred especially with untreated sludge and hydrochar from

lower treatment temperatures. Shrinkage however decreased considerably with an increase in

hydrothermal treatment temperature, which also increased effective diffusivity of hydrochar at lower

moisture contents (Fig. 3f). Previously Zhao et al. (2014a) investigated the drying properties of

hydrothermally treated paper sludge and reported effective diffusivities in the range 1.26-1.71 ∙ 10-9

m2 s-1 for hydrochar treated at 180-260 °C for 30 minutes. The authors used drying temperatures of 30

°C, did not take shrinkage into account and used a linear model for describing the evolution of

moisture ratio as a function of drying time. However, Bennamoun et al. (2013) reported a diffusivity

difference of 106-134% if shrinkage is not taken into account during laboratory drying experiments

on municipal sludge.

The attained regression model indicated that hydrothermal treatment temperature was 1.6 times as

important as transformed retention time in controlling the mean moisture diffusivity of hydrochar

(Fig. S1). As the experimental variables were coded and range-scaled their effects could be compared

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

13

within the original design range. In addition, the dummy variables used for describing the effect of

acid and base additions suggested that the use of NaOH or HCl did not significantly affect the mean

moisture diffusivity of hydrochar (p > 0.05). Previously the effect of process conditions have mainly

been investigated in terms of solid fuel properties and yields of attained hydrochar. Although it is

currently well known that temperature mainly governs biomass decomposition in hydrothermal

media, retention time has also been found significant for the fuel properties of hydrochar produced

from algae and municipal and industrial sludge (Danso-Boateng et al., 2015; Heilmann et al., 2011;

Mäkelä et al., 2015; Xu et al., 2013). In addition, Lynam et al. (2011) found that organic acid

additions enhanced cellulose dissolution and respectively decreased hydrochar yield during

hydrothermal treatment of lignocellulosic biomass. Furthermore, Lu et al. (2014) reported statistically

different solid yields of hydrothermally treated cellulose in NaOH and HCl due to changes in reaction

kinetics. In this work the mean diffusivities of hydrochar were higher with the use of H2O than with

NaOH or HCl, but the overall differences were not statistically significant (Fig. S1). Liquid pH values

measured after the hydrothermal experiments were in the range 4.9-5.5 and correlated only with

treatment temperature (p < 0.01) and retention time (p < 0.05). This suggests that the strength of used

additives were not sufficient to significantly change reaction conditions and were neutralized likely

due to the formation of organic acids during biomass decomposition under hydrothermal conditions

(Berge et al., 2011; Weiner et al., 2014). Equivalent hydroxide and hydronium ion concentrations

from NaOH and HCl have previously been used for simulating different pH environments in

municipal and industrial waste streams (Lu et al., 2014). In addition, equivalent hydroxide

concentrations from KOH have been reported to lead to decreased surface area and pore volume of

hydrothermally treated wheat straw (Reza et al., 2015) which could have practical implications for the

drying behavior of hydrochar. Besides evaluating the effect of process conditions the attained model

coefficients enable prediction of novel observations, which is very useful for illustrating the behavior

of mean diffusivity of hydrochar within the design range. Respective contours were hence calculated

based on hydrothermal treatment temperature and retention time in NaOH, H2O and HCl and are

shown in Fig. 4. The linear increase in mean diffusivity as a function reaction temperature and

retention time could easily be observed and was more pronounced with NaOH or H2O compared to

HCl. As an example, hydrothermal treatment at 220 °C with a retention time of 1.5 h in H2O would be

expected to increase the mean effective diffusivity of sludge to 18.2 ± 5.80 ∙ 10-9 m2 s-1 with 95%

certainty. According to our better knowledge this is the first time the effect and statistical significance

of hydrothermal treatment conditions on the mean moisture diffusivity and drying behavior of sludge

have been reported. Once these effects are known, the mean diffusivity of sludge can be predicted

within the design thus allowing control of respective drying properties based on hydrothermal

treatment conditions. This information is vital for optimizing hydrothermal treatment and subsequent

drying processes for this feedstock.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

14

Fig. 4: Predicted mean effective moisture diffusivity (10-9 m2 s-1) of hydrochar as a function of

reaction temperature (°C) and retention time (h) in NaOH, H2O and HCl.

Conclusions

Hydrothermal treatment enhanced the drying properties and increased the effective moisture

diffusivity of sludge. Drying of untreated sludge was governed by the falling rate period where drying

flux decreased continuously as a function of sludge moisture content due to heat and mass transfer

limitations and sample shrinkage. Hydrothermal treatment increased the drying flux of sludge

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

15

hydrochar and decreased the effect of internal heat and mass transfer limitations and sample shrinkage

especially at higher treatment temperatures. The determined effective moisture diffusivities increased

as a function of decreasing moisture content and a statistically significant difference was found

between the mean diffusivity of untreated sludge and sludge hydrochar. The attained regression model

indicated that treatment temperature controlled the mean moisture diffusivity of hydrochar as the

effects of NaOH and HCl were found statistically insignificant. The presented results enabled

prediction of sludge drying properties through mean moisture diffusivity based on hydrothermal

treatment conditions.

Acknowledgements

The contributions of Markus Segerström and Gunnar Kalén from the Swedish University of

Agricultural Sciences in sample processing are deeply appreciated. This work was performed in

cooperation with SCA Obbola AB with partial funding from SP Processum AB.

References

Alatalo, S.-M., Repo, E., Mäkilä, E., Salonen, J., Vakkilainen, E., Sillanpää, M., 2013. Adsorption

behavior of hydrothermally treated municipal sludge & pulp and paper industry sludge. Bioresource

Technology 147, 71-76. doi: 10.1016/j.biortech.2013.08.034.

Bennamoun, L., Crine, M., Léonard, A., 2013. Convective drying of wastewater sludge: introduction

of shrinkage effect in mathematical modeling. Drying Technology 31, 643-654. doi:

10.1080/07373937.2012.752743.

Berge, N.D., Ro, K.S., Mao, J., Flora, J.R.V., Chappell, M.A., Bae, S., 2011. Hydrothermal

carbonization of municipal waste streams. Environmental Science and Technology 45, 5696-5703.

doi: 10.1021/es2004528.

Crank, J. The Mathematics of Diffusion (2nd ed.). Oxford University Press: London, 1975, 414 pp.

Danso-Boateng, E., Shama, G., Wheatley, A.D., Martin, S.J., Holdich, R.G., 2015. Hydrothermal

carbonisation of sewage sludge: effect of process conditions on product characteristics and methane

production. Bioresource Technology 177, 318-327. doi: 10.1016/j.biortech.2014.11.096.

Dawei, M., Zili, J., Yoshikawa, K., Hongyan, M., 2012. The effect of operation parameters on the

hydrothermal drying treatment. Renewable Energy 42, 90-94. doi: 10.1016/j.renene.2011.09.011.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

16

Fraikin, L., Contribution à l'étude du séchage convectif de boues de station d'épuration et des

émissions gazeuses associées, University of Liège, Department of Chemical Engineering, 2012.

Fyhr, C., Kemp, I.C., 1998. Comparison of different drying kinetics models for single particles.

Drying Technology 16, 1339-1369. doi: 10.1080/07373939808917465.

Gerlach, R.W., Dobb, D.E., Raab, G.A., Nocerino, J.M., 2002. Gy sampling theory in environmental

studies. 1. Assessing soil splitting protocols. Journal of Chemometrics 16, 321-328. doi:

10.1002/cem.705.

Gómez-de la Cruz, F.J., Palomar-Carnicero, J.M., Casanova-Peláez, P.J., Cruz-Peragón, F., 2015.

Experimental determination of effective moisture diffusivity during the drying of clean olive stone:

dependence of temperature, moisture content and sample thickness. Fuel Processing Technology 137,

320-326. doi: 10.1016/j.fuproc.2015.03.018.

Heilmann, S.M., Jader, L.R., Sadowsky, M.J., Schendel, F.J., von Keitz, M.G., Valentas, K.J., 2011.

Hydrothermal carbonization of distiller's grains. Biomass and Bioenergy 35, 2526-2533. doi:

10.1016/j.biombioe.2011.02.022.

Kruse, A., Funke, A., Titirici, M.-M., 2013. Hydrothermal conversion of biomass to fuels and

energetic materials. Current Opinion in Chemical Biology 17, 515-521. doi:

10.1016/j.cbpa.2013.05.004.

Léonard, A., Blacher, S., Marchot, P., Crine, M., 2002. Use of X-ray microtomography to follow the

convective heat drying of wastewater sludges. Drying Technology 20, 1053-1069. doi: 10.1081/DRT-

120004013.

Léonard, A., Blacher, S., Marchot, P., Pirard, J.P., Crine, M., 2004. Measurement of shrinkage and

cracks associated to convective drying of soft materials by X-ray microtomography. Drying

Technology 22, 1695-1708. doi: 10.1081/DRT-200025629.

Léonard, A., Meneses, E., Le Trong, E., Salmon, T., Marchot, P., Toye, D., Crine, M., 2008.

Influence of back mixing on the convective drying of residual sludges in a fixed bed. Water Research

42, 2671-2677. doi: 10.1016/j.watres.2008.01.020.

Li, J., Bennamoun, L., Fraikin, L., Salmon, T., Toye, D., Schreinemachers, R., Léonard, A., 2014.

Analysis of the shrinkage effect on mass transfer during convective drying of sawdust/sludge

mixtures. Drying Technology 32, 1706-1717. doi: 10.1080/07373937.2014.924136.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

17

Libra, J.A., Ro, K.S., Kammann, C., Funke, A., Berge, N.D., Neubauer, Y., Titirici, M.-M., Fühner,

C., Bens, O., Kern, J., Emmerich, K.-H., 2011. Hydrothermal carbonization of biomass residuals: a

comparative review of the chemistry, processes and applications of wet and dry pyrolysis. Biofuels 2,

71-106. doi: 10.4155/BFS.10.81.

Lu, X., Flora, J.R.V., Berge, N.D., 2014. Influence of process water quality on hydrothermal

carbonization of cellulose. Bioresource Technology 154, 229-239. doi:

10.1016/j.biortech.2013.11.069.

Lynam, J.G., Coronella, C.J., Yan, W., Reza, M.T., Vasquez, V.R., 2011. Acetic acid and lithium

chloride effects on hydrothermal carbonization of lignocellulosic biomass. Bioresource Technology

102, 6192-6199. doi: 10.1016/j.biortech.2011.02.035.

Mahmood, T., Elliott, A., 2006. A review of secondary sludge reduction technologies for the pulp and

paper industry. Water Research 40, 2093-2112. doi: 10.1016/j.watres.2006.04.001.

Mäkelä, M., Benavente, V., Fullana, A., 2015. Hydrothermal carbonization of lignocellulosic

biomass: effect of process conditions on hydrochar properties. Applied Energy 155, 576-584. doi:

10.1016/j.apenergy.2015.06.022.

Mowla, D., Tran, H.N., Grant Allen, D., 2013. A review of the properties of biosludge and its

relevance to enhanced dewatering processes. Biomass and Bioenergy 58, 365-378. doi:

10.1016/j.biombioe.2013.09.002.

Mujumdar, A.S., Principles, Classification and Selection of Dryers. In: Handbook of Industrial Drying

(3rd ed.), A. S. Mujumdar (Eds.), CRC Press: Boca Raton; 2007, pp. 4-32.

Myers, R.H., Montgomery, D.C., Anderson-Cook, C.M., Experimental Designs for Fitting Response

Surfaces - II. In: Response Surface Methodology: Process and Product Optimization Using Designed

Experiments (3rd ed.), R. H. Myers, D. C. Montgomery and C. M. Anderson-Cook (Eds.), John Wiley

& Sons, Inc.: Hoboken; 2009, pp. 349-416.

Pacheco-Aguirre, F.M., Ladrón-González, A., Ruiz-Espinosa H, García-Alvarado, M.A., Ruiz-López,

I.I., 2014. A method to estimate anisotropic diffusion coefficients in cylindrical solids: application to

the drying of carrot. Journal of Food Engineering 125, 24-33. doi: 10.1016/j.jfoodeng.2013.10.015.

Peterson, A.A., Vogel, F., Lachance, R.P., Fröling, M., Antal Jr., M.M., Tester, J.W., 2008.

Thermochemical biofuel production in hydrothermal media: a review of sub- and supercritical water

technologies. Energy & Environmental Science 1, 32-65. doi: 10.1039/b810100k.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

18

Reza, M.T., Rottler, E., Herklotz, L., Wirth, B., 2015. Hydrothermal carbonization (HTC) of wheat

straw: influence of feedwater pH prepared by acetic acid and potassium hydroxide. Bioresource

Technology 182, 336-344. doi: 10.1016/j.biortech.2015.02.024.

Sevilla, M., Fuertes, A.B., 2009. The production of carbon materials by hydrothermal carbonization

of cellulose. Carbon 47, 2281-2289. doi: 10.1016/j.carbon.2009.04.026.

Stoica, A., Sandberg, M., Holby, O., 2009. Energy use and recovery strategies within wastewater

treatment and sludge handling at pulp and paper mills. Bioresource Technology 100, 3497-3505. doi:

10.1016/j.biortech.2009.02.041.

Titirici, M.-M., Antonietti, M., 2010. Chemistry and material options of sustainable carbon materials

made by hydrothermal carbonization. Chemical Society Reviews 39, 103-116. doi:

10.1039/b819318p.

Vaxelaire, J., Cézac, P., 2004. Moisture distribution in activated sludges: a review. Water Research

38, 2215-2230. doi: 10.1016/j.watres.2004.02.021.

Wang, L., Zhang, L., Li, A., 2014. Hydrothermal treatment coupled with mechanical expression at

increased temperature for excess sludge dewatering: influence of operating conditions and process

energetics. Water Research 65, 85-97. doi: 10.1016/j.watres.2014.07.020.

Weiner, B., Poerschmann, J., Wedwitschka, H., Koehler, R., Kopinke, F.-D., 2014. Influence of

process water reuse on the hydrothermal carbonization of paper. ACS Sustainable Chemistry &

Engineering 2, 2165-2171. doi: 10.1021/sc500348v.

Xu, Q., Qian, Q., Quek, A., Ai, N., Zeng, G., Wang, J., 2013. Hydrothermal carbonization of

macroalgae and the effects of experimental parameters on the properties of hydrochars. ACS

Sustainable Chemistry & Engineering 1, 1092-1101. doi: 10.1021/sc400118f.

Yoshikawa, K., Prawisudha, P., Hydrothermal Treatment of Municipal Solid Waste for Producing

Solid Fuel. In: Application of Hydrothermal Reactions to Biomass Conversion (1st ed.), F. Jin (Eds.),

Springer: Heidelberg; 2014a, pp. 355-383.

Yoshikawa, K., Prawisudha, P., Sewage Sludge Treatment by Hydrothermal Process for Producing

Solid Fuel. In: Application of Hydrothermal Reactions to Biomass Conversion (1st ed.), F. Jin (Eds.),

Springer: Heidelberg; 2014b, pp. 385-409.

An accepted version of Mäkelä et al., Water Res 91 (2016) 11-18, available at: http://dx.doi.org/10.1016/j.watres.2015.12.043.

19

Zhao, P., Ge, S., Ma, D., Areeprasert, C., Yoshikawa, K., 2014a. Effect of hydrothermal pretreatment

on convective drying characteristics of paper sludge. ACS Sustainable Chemistry & Engineering 2,

665-671. doi: 10.1021/sc4003505.

Zhao, P., Shen, G., Ge, S., Chen, Z., Yoshikawa, K., 2014b. Clean solid biofuel production from high

moisture content waste biomass employing hydrothermal treatment. Applied Energy 131, 345-367.

doi: 10.1016/j.apenergy.2014.06.038.

The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.